branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.3.14;
Update the devlist2h.awk script to track the maximum lengths of vendor
and product strings, and report the max values at end of the run.

Update the Makefiles.{pci,usb,hdaudio}devs to point users at the places
which might need to be updated if the maximum lengths get larger.

Since this commit makes no changes to the generated files, we don't
need to regenerate them now.

branches: 1.2.2;
fix path to devlist2h

branches: 1.1.2;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

hdaudio: Obsolete HDAUDIO_ENABLE_DISPLAYPORT and HDAUDIO_ENABLE_HDMI

Enable display port and HDMI devices by default rather than making this
a compile time option.

Replace the HDAUDIO_32BIT_ACCESS option with a runtime option. Set it for
tegra_hdaudio, but not for the pci attachment. Add hdaudio@pci to GENERIC64

Merge isaki-audio2 branch, the overhaul of audio subsystem.
- Interrupt-oriented system rather than thread-oriented.
- Improve stability, quality and performance.
- Split playback and record cleanly. Improve halfduplex support.
- Many bugs are fixed including deadlocks, resource leaks, abuses, etc.
- Simplify audio filter mechanism. The encoding/channels/frequency
conversions are completely handled in the upper layer. So the hard-
ware driver only converts its hardware encoding (if necessary).
- audio_hw_if changes:
- Obsoletes query_encoding and add query_format instead.
- Obsoletes set_params and add set_format instead.
- Remove drain, setfd, mappage.
- The call sequences are changed.
- ioctl AUDIO_GETFD/SETFD, AUDIO_GETCHAN/SETCHAN are obsoleted.
- ioctl AUDIO_{QUERY,GET,SET}FORMAT are introduced.
- cleanup config attributes: au*conv and mulaw.
- All hardware drivers should follow it (I've done as much as possible).

Some file paths are changed:
- dev/audio.c -> dev/audio/audio.c (rewritten)
- dev/audiovar.h -> dev/audio/audiovar.h
- dev/audio_dai.h -> dev/audio/audio_dai.h
- dev/audio_if.h -> dev/audio/audio_if.h
- dev/audiobell.c -> dev/audio/audiobell.c
- dev/audiobellvar.h -> dev/audio/audiobellvar.h
- dev/mulaw.[ch] -> dev/audio/mulaw.[ch] + dev/audio/alaw.c

branches: 1.3.16; 1.3.20; 1.3.24;
add HDAUDIO_32BIT_ACCESS option, which wraps mmio read/writes to only use 32-bit accesses

branches: 1.2.2;
replace hdaudio_config.h with defflags

Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

Remove the call to kthread_join in hdafg_detach. The thread exits when sc_jack_polling is
cleared and the thread is not created with the KTHREAD_MUSTJOIN flag set.

This fixes a hard hang in kthrwait when the hdaudioctl(1) command is used to install a custom
configuration.

hdaudio(4): Do not expose non-functional volume controls.

The old code supported "volume without mute", but not "mute
without volume", which are widely exposed by hdaudio devices.
This resulted in redundant volume controls with delta=256
(not allowing them to be set).

While here, classify PC speaker and phone out controls as
output controls.

"ship it" jmcneill@

PR kern/59712 hdaudio driver exposes nonexistent volume controls

branches: 1.32.4;
hdaudio(4): KNF, no functional change intended

hdaudio(4): Tighten hdafg_freem.

The fix for PR kern/57890 in hdafg.c 1.29 included more conditionals
than needed, covering cases that can't happen: if allocm returns
null, audio(4) does not call freem; and if hdafg_allocm returns
nonnull, then what audio(4) passes to hdafg_freem had better be
either the playback or the capture buffer address.

hdafg(4): Do hotplug detection in kthread, not callout.

This can sometimes take a while (~1ms), and the logic to suspend the
callout on device suspend/resume was racy (PR kern/57322).

XXX pullup-8
XXX pullup-9
XXX pullup-10

ignore NULL for hdaudio_audiodev in hdafg_freem
also guard against NULL in ad_{capture,playback}
observed during shutdown on Lenovo W510

branches: 1.28.4;
fix various typos in comments, docs and log messages.

hdaudio: Obsolete HDAUDIO_ENABLE_DISPLAYPORT and HDAUDIO_ENABLE_HDMI

Enable display port and HDMI devices by default rather than making this
a compile time option.

Safe vendor/product for reporting. Avoids repeated loading/unloading
of hdaudioverbose module.

One build break fixed, another appears...

Use the correct pointer, not the one used in the other piece of
code that changed at the same time..

If the step is 0, use 1 instead of always incrementing step by one to avoid
zerodivide (from RVP)

Update for proplib(3) API changes.

Make round_blocksize satisfy all of
- restrictions that existed before merging isaki-audio2 branch.
- better support for 6 channels hardware.
- audio layer's requirement.
This may help PR kern/54474.

branches: 1.21.4;
Enumerating probably always starts from sc->sc_startnode.

Only dump ELD info if HDAFG_HDMI_DEBUG is defined

Make sure we have at least one of playback or capture streams setup before
attaching the audio layer.

branches: 1.18.2; 1.18.4;
Clean get_props().
- Make get_props() return AUDIO_PROP_{PLAYBACK,CAPTURE} properly.
This eliminates need for audio.c to take care of such (old)
drivers which don't return both of PLAYBACK and CAPTURE.
- All get_props() doesn't need to return AUDIO_PROP_MMAP.
It is handled in the audio layer now.

Merge isaki-audio2 branch, the overhaul of audio subsystem.
- Interrupt-oriented system rather than thread-oriented.
- Improve stability, quality and performance.
- Split playback and record cleanly. Improve halfduplex support.
- Many bugs are fixed including deadlocks, resource leaks, abuses, etc.
- Simplify audio filter mechanism. The encoding/channels/frequency
conversions are completely handled in the upper layer. So the hard-
ware driver only converts its hardware encoding (if necessary).
- audio_hw_if changes:
- Obsoletes query_encoding and add query_format instead.
- Obsoletes set_params and add set_format instead.
- Remove drain, setfd, mappage.
- The call sequences are changed.
- ioctl AUDIO_GETFD/SETFD, AUDIO_GETCHAN/SETCHAN are obsoleted.
- ioctl AUDIO_{QUERY,GET,SET}FORMAT are introduced.
- cleanup config attributes: au*conv and mulaw.
- All hardware drivers should follow it (I've done as much as possible).

Some file paths are changed:
- dev/audio.c -> dev/audio/audio.c (rewritten)
- dev/audiovar.h -> dev/audio/audiovar.h
- dev/audio_dai.h -> dev/audio/audio_dai.h
- dev/audio_if.h -> dev/audio/audio_if.h
- dev/audiobell.c -> dev/audio/audiobell.c
- dev/audiobellvar.h -> dev/audio/audiobellvar.h
- dev/mulaw.[ch] -> dev/audio/mulaw.[ch] + dev/audio/alaw.c

branches: 1.16.2;
Fix hdaudio device configuration

When disabling unassociated devices, we have a special handling
for pins of type COP_AWCAP_TYPE_PIN_COMPLEX, but it came after code
that may disable any pins, including the ones that should be handled
as COP_AWCAP_TYPE_PIN_COMPLEX.

The result was that hdaudio could fail to detect some devices. We
fix the situation by making sure that COP_AWCAP_TYPE_PIN_COMPLEX
pins always get their specific handling.

The change makes the built-in speaker work on Dell OptiPlex 5060

branches: 1.15.2; 1.15.4;
Remove dead code.

From David Binderman in PR kern/53029

Fix off-by-one when calling snprintf(9) in hdafg_getdev()

This is actually harmless, since:
- the offset is too short rather than too long (no overflow)
- the struct audio_device comes from userland (no information leak)

"looks good to me" nat@

branches: 1.13.2;
put all the ELD debugging messages under #ifdef HDAFG_HDMI_DEBUG.
this silences a frequent and largely useless series of messages
in my dmesg. ok jmcneill.

branches: 1.12.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

remove redundant assignments
nchan is overwritten before it is used in all cases. so is i.

from clang static analyzer
ok riastradh

branches: 1.10.2;
Replace numeric magic-number constant with something a bit more meaningful.

branches: 1.9.2;
disable data island packet transmission on Tegra124 HDMI codec

mute the mixer with that has input children

Remove trailing and stray whitespace.

more complex version to determine what is an input.

be more selective and mute only mic's. Suggested by jmcneill

Startup with the mic muted to avoid feedback and spying.

- Support multi-channel (> 2) output to HDMI sinks.
- Fix a bug in hdafg_set_params that could program converters using the
wrong audio_params_t if auconv is in use
- Force Tegra124 HDMI codec to a fixed rate 44.1kHz

branches: 1.2.2;
replace hdaudio_config.h with defflags

Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.3.2;
In rev 1.0a of the Intel High Definition Audio Spec:

https://www.intel.com/content/www/us/en/standards/
high-definition-audio-specification.html

page 186 shows the layout of the baseline block of the ELD (EDID Like Data)
struct - and allows a reserved (effectively padding) area at the end of the
struct. This is required to keep the struct an even number of words long
(size measured in units of 32 bits) while allowing for a variable length
monitor name, followed by a variable number of 3 byte structs - the
combination of which is not likely to be a multiple of 4.

Code here assumed that there was no padding, and objected to the ELD
format if any padding bytes existed (hdafg_dd_parse_info() would return
EINVAL) causing a "failed to parse ELD data" message (if HDAFG_HDMI_DEBUG
is defined) from hdafg_assoc_dump_dd() making it difficult (or at least
confusing) to debug HDMI related audio issues (hdafg_assoc_dump_dd would
not print most of the data it is expected to print) although this would
most likely have no effect on actual operations.

Change a test from a != to < (there must be enough data, not exactly the
amount needed) for the EINVAL. As a consequence, the length after the
SAD data is parsed (the 3 byte structs) is no longer required to be 0,
so remove the KASSERT() (previously it was just useless, the code guaranteed
a 0 value, now it is incorrect.) While here also change a related
diagnostic message to be slightly more informative as to what is being shown.

OK jmcneill@

put all the ELD debugging messages under #ifdef HDAFG_HDMI_DEBUG.
this silences a frequent and largely useless series of messages
in my dmesg. ok jmcneill.

branches: 1.1.2; 1.1.12;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

fix various typos in comments.

driver(9): devsw_detach never fails. Make it return void.

Prune a whole lotta dead branches as a result of this. (Some logic
calling this is also wrong for other reasons; devsw_detach is final
-- you should never have any reason to decide to roll it back. To be
cleaned up in subsequent commits...)

XXX kernel ABI change to devsw_detach signature requires bump

Merge thorpej-cfargs2.

branches: 1.15.2;
Rework the xxxVERBOSE option to share the common module-hook-based
verbose mechanism with MIIVERBOSE. This reduces some duplicated code
and allows us to once again permit auto-unload of MIIVERBOSE.

Change details:
* Update dev/devlist2h.awk to accomodate miidevs, including generation
of MII_STR_oui_model definitions and use of oui and model rather than
vendor and product. This also changes the compressed data in the
xxxdevs_data.h files to uint32_t (since mii oui's are up to 6 hex
digits long)
* Update a couple of phy drivers to use new calls to get verbose data
* Regen all of the xxxdevs{,_data}.h files (separate commit, coming
very soon)
* Update mii/mii_verbose.[ch] and mii/mii_physubr.c to use the various
DEV_VERBOSE_xxx macros
* Update the pci, usb, and hdaudio code as needed, to #include the
xxxdevs.h files (in order to get the proper printf format strings)
* Since dev/dev_verbose.c now uses non-literal printf format strings,
(to deal with the vendor/product vs oui/model issue), we need to
make sure it gets compiled with -Wno-error=format-nonliteral, even
in userland's libpci and librumpdev!
* Bump kernel version for the change in module interfaces

Welcome to 9.99.86!

XXX It might be useful in the future to extend the MII_STR_oui_model
XXX definitions to PCI as well (and perhaps USB and HDAUDIO). This
XXX would allow for a single centralized location for the products'
XXX descriptions, rather than being dispersed among individual
XXX drivers' xxx_match tables.

branches: 1.14.2;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.13.2;
If the Subordinate Node Count returns 0 nodes, complain and return instead
of trying to kmem_zalloc 0 bytes later on.

Initialize DMA memory when allocating it. Should fix PR# 51734 (hdaudio
"RIRB timeout" on boot).

branches: 1.11.2;
Update for proplib(3) API changes.

Read GCAP and version regs after taking the controller out of reset. Fixes
stream counts on my Radeon HD 7850.

branches: 1.9.2; 1.9.4;
Even though the spec says SRST should read back as 1 after being set, this
does not appear to be the case with QEMU when using single byte accesses.

Instead of printing an error and giving up, continue and try to leave
stream reset state even if setting SRST times out.

branches: 1.8.2; 1.8.6;
Always go through RIRB startup process, initialize RIRB interrupt count
register, and ack RIRBs as we process them in polling mode.

XXX pullup

Always access CORBCTL using hda_read1/hda_write1 (it is an 8-bit wide reg).
Reported by Michal Necasek.

Enter link reset even if GCTL says we are already in reset state. Fixes
hdaudio codec detection under VirtualBox on a "cold" boot.

XXX pullup

Move the hdaudio_cd out of the hdaudio_pci module and into the main
hdaudio module. Rearrange the inter-module dependencies accordingly.

Should resolve (my) PR kern/49836

XXX Pull-ups needed for netbsd-8 branch

branches: 1.4.8; 1.4.10;
get rid of HDAUDIO_FLAG_NO_STREAM_RESET hack

Skip stream reset if HDAUDIO_FLAG_NO_STREAM_RESET flag is set.

Initialize dma.dma_sizereg to avoid copying uninitialized data later.

This field is not actually used here -- not clear it should be in
the hdaudio_dma structure at all.

Noted by Coverity, CID 979553.

branches: 1.1.2;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

replace hdaudio_config.h with defflags

Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

hdaudio(4): Rename "record" mixer control to "imonitor".

This more accurately represents what the control actually does.

If you happen to have a configuration file that sets recording
monitor mixer levels, you'll need to modify it. Therefore this
isn't suitable for pulling up to existing stable branches.

"go nuts" jmcneill@

branches: 1.1.2; 1.1.18; 1.1.62;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

hdaudio: Add NVIDIA GK208, REALTEK ALC897

add realtek ALC256.

Add Realtek ALC233

branches: 1.10.2;
Add Cirrus Logic CS4206

branches: 1.9.2;
Add VIA VX900 HDMI.

Add KX-6000 (ZX-E) HDA codecs.

Add Realtek ALC671 to hdaudiodevs, seen on Fujitsu Futro S720 thin client

branches: 1.6.4;
hdaudiodevs: Add some ~new Intel devices

hdaudiodevs: Add Realtek ALC292

Add Realtek ALC293

Make sure that all vendor strings are shorter than 16 (MAX_AUDIO_DEV_LEN)
characters as the values are used in the AUDIO_GETDEV ioctl.

branches: 1.2.16; 1.2.20; 1.2.26; 1.2.28;
add Tegra124 HDMI

branches: 1.1.2;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

regen

regen.

regen

branches: 1.11.2;
Regen

branches: 1.10.2;
Regen.

regen.

regen

branches: 1.7.4;
regen

Regen for new sys/dev/devlist2h.awk

branches: 1.5.6;
hdaudiodevs: regen

regen

regen

branches: 1.2.16; 1.2.20; 1.2.26; 1.2.28;
regen

branches: 1.1.2;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

regen

regen.

regen

branches: 1.11.2;
Regen

branches: 1.10.2;
Regen.

regen.

regen

branches: 1.7.4;
regen

Regen for new sys/dev/devlist2h.awk

branches: 1.5.6;
hdaudiodevs: regen

regen

regen

branches: 1.2.16; 1.2.20; 1.2.26; 1.2.28;
regen

branches: 1.1.2;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

hdaudio: pci: PCIe config space fixes and Intel PCH snoop support

The HD audio specification does not cover PCI config space, and this
driver was unconditionally writing to a vendor specific register. Reduce
scope of config space accesses based on PCI IDs.

With this cleaned up, add support for Intel PCH devices which require
some additional vendor specific configuration to bypass no snoop mode.

s/1 << 31/1u << 31/.

XXX someone could __BIT() etc this file.

branches: 1.2.2; 1.2.4; 1.2.6;
Add #includes to make this compile.

<sys/cdefs.h> for __packed
<sys/types.h> for uint32_t

branches: 1.1.2;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

Replace the HDAUDIO_32BIT_ACCESS option with a runtime option. Set it for
tegra_hdaudio, but not for the pci attachment. Add hdaudio@pci to GENERIC64

Merge isaki-audio2 branch, the overhaul of audio subsystem.
- Interrupt-oriented system rather than thread-oriented.
- Improve stability, quality and performance.
- Split playback and record cleanly. Improve halfduplex support.
- Many bugs are fixed including deadlocks, resource leaks, abuses, etc.
- Simplify audio filter mechanism. The encoding/channels/frequency
conversions are completely handled in the upper layer. So the hard-
ware driver only converts its hardware encoding (if necessary).
- audio_hw_if changes:
- Obsoletes query_encoding and add query_format instead.
- Obsoletes set_params and add set_format instead.
- Remove drain, setfd, mappage.
- The call sequences are changed.
- ioctl AUDIO_GETFD/SETFD, AUDIO_GETCHAN/SETCHAN are obsoleted.
- ioctl AUDIO_{QUERY,GET,SET}FORMAT are introduced.
- cleanup config attributes: au*conv and mulaw.
- All hardware drivers should follow it (I've done as much as possible).

Some file paths are changed:
- dev/audio.c -> dev/audio/audio.c (rewritten)
- dev/audiovar.h -> dev/audio/audiovar.h
- dev/audio_dai.h -> dev/audio/audio_dai.h
- dev/audio_if.h -> dev/audio/audio_if.h
- dev/audiobell.c -> dev/audio/audiobell.c
- dev/audiobellvar.h -> dev/audio/audiobellvar.h
- dev/mulaw.[ch] -> dev/audio/mulaw.[ch] + dev/audio/alaw.c

branches: 1.6.2; 1.6.6;
s/static inline/static __inline/g for consistency.

branches: 1.5.16; 1.5.18;
get rid of HDAUDIO_FLAG_NO_STREAM_RESET hack

Skip stream reset if HDAUDIO_FLAG_NO_STREAM_RESET flag is set.

add HDAUDIO_32BIT_ACCESS option, which wraps mmio read/writes to only use 32-bit accesses

branches: 1.2.2;
replace hdaudio_config.h with defflags

Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.

branches: 1.1.2; 1.1.18;
Split hdaudio and PCI attach glue. Even though the driver was written with
this separation in mind, all of the code lived in sys/dev/pci/hdaudio. Move
common parts to sys/dev/hdaudio and PCI attach glue to sys/dev/pci.